Re: [gnugo-devel] Owl persistent cache

[Arend Bayer]

On Thu, 28 Feb 2002, Daniel Bump wrote:

>
> Inge wrote:
>
> > After computing the result, it recalculated it again using the
> > contents of the hash table.  Thus many blind alleys were never
> > searched, and the active area turned out much smaller, sometimes only
> > half as big as before.
>
> This seems drastic since then the code is run twice.
This method would look doable if the second run was some magnitudes faster.
It would be automatically faster because in WIN situations, we know the correct
answer at once. Unfortunately, in LOOSE situations, we would have to go
through the move generation to produce the moves tested, which is the
most expensive part and which requires all the other expensive work of
updating the owl data.

However, we could do the following method: in a LOOSE situation (which can
be recognized from the hash table), we just try out all moves close to the
dragon to be attacked, until we find MAX_MOVES moves that result in a
position stored in the hash table as WIN for the opponent.

This would probably be fast enough unless the hash table lookup is too
time consuming.

Arend